Pedro, you use just utility matrix.
Why don't you use movie features, user features?
because the utility matrix gives a more accurate prediction?
or you just looked at an example, without touching movie features, user features?

Environment: Python 3.5.4rc1
when I build the topic model in jupyter notebook, python always throw .local/lib/python3.5/site packages/gensim/models/ldamodel.py:802: RuntimeWarning: divide by zero encountered in log, I don't know why np.log(0) happens.

It seems that the codes in ch09 are not correct. Some preprocess steps are missing?

Hi Luis, I'm starting to learning machine learning with scikit, I'm from Brazil, is it possible to show a example of prediction for the sentiment analisys at the chapter6 after have the model trained ?
I want to check with a single tweet how much is he positive or negative, base on the == Pos vs. neg == trained model. Can you show a example please.

thanks

Hi! Thank you for create the book, I'm using it and I find it awesome! But I'm getting a problem with ch05/so_xml_to_tsv.py, at some point the execution gets killed because the huge amount of memory that it's using (with a 16GB RAM pc). I'll try to see which objects are creating this problem and maybe fix the issue, so, I'll keep you informed.

When trying to load the data, it returns the exceptions case.

@luispedro This code on line 88 of Ch01/analyze_webstats.py (and page 22 in the book)

xa = x[:inflection]
ya = y[:inflection]
xb = x[inflection:]
yb = y[inflection:]

Since you trimmed the list on line 22, so the values corresponding to some indexes have been removed.
For example, x[2] has been trimmed out in line 22, and more, this leads to left shift in the values contained in the list. What I mean by that is that now that for x[inflection-1] will contain a value which is actually larger than inflection. This division would have worked fine for the raw untrimmed array but not after it's been trimmed.

Please fix this in the next edition or if I'm wrong then please make me understand how is this correct.

According to this (http://blog.stackexchange.com/category/cc-wiki-dump/) the websites provided for the data in the book are no longer valid. The blog claims that the data is now at the Internet Archive. However, I cannot seem to locate the data necessary for chapter 5.

Hi Luis,

I have just started following the book for getting started with machine learning. However, the below code in threshold.py yields a boolean value and not a numpy array on which mean can be calculated.

Good day!
I have a mistake in BuildingMachineLearningSystemsWithPython/ch03/plot_kmeans_example.py

`import utils
CHART_DIR = 'D:\Python'
from utils import CHART_DIR

ImportError: cannot import name 'CHART_DIR'`

what did i do wrong?

I see the following error when running ch04/blei_lda.py

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-2-5511c9e27cd2> in <module>()
     37 for ti in range(model.num_topics):
     38     words = model.show_topic(ti, 64)
---> 39     tf = sum(f for f, w in words)
     40     with open('topics.txt', 'w') as output:
     41         output.write('\n'.join('{}:{}'.format(w, int(1000. * f / tf)) for f, w in words))

TypeError: unsupported operand type(s) for +: 'int' and 'str'

I think, since the frequency, f, and word, w, are the opposite on words, did you want to write w, f in words instead, as in the following?

     39     tf = sum(f for w, f in words)
     40     with open('topics.txt', 'w') as output:
     41         output.write('\n'.join('{}:{}'.format(w, int(1000. * f / tf)) for w, f in words))

BuildingMachineLearningSystemsWithPython/ch05/so_xml_to_tsv.py

in the parsexml function, I had to make the following changes to have it working:

def parsexml(filename):
    global num_questions, num_answers

    counter = 0

    it = iter(etree.iterparse(filename, events=('start',)))

    event, root = next(it)  # get posts element

    for event, elem in it:

When I run the utils.py file, the following text appears:
"Please set GENRE_DIR and TEST_DIR in utils.py"

leaving the comment: "Process finished with exit code 1"
and I have no idea what to do and how to do. Help me please!

Note: I have sample .wav files in a folder "genres" inside the same directory where my python project exists. That folder further contains subfolders for samples in each music genre